1 node04
45124 test stress.s user2 R 4:24 1 node04
From: slurm-users on behalf of Walls,
Mitchell
Sent: Friday, February 18, 2022 10:24 AM
To: slurm-users@lists.schedmd.com
Subject: Re: [slurm-users] Suspend QOS help
Time
test stress-s user1 S 2:00 1 node04
#partition
PartitionName=testDefault=NO Nodes=node04 OverSubscribe=FORCE:1
MaxTime=30-00:00:00 AllowQos=general,suspend
From: slurm-users on behalf of Walls,
Mitchell
Sent: Friday, F
=suspend AllowQos=suspend
From: Walls, Mitchell
Sent: Friday, February 18, 2022 9:54 AM
To: slurm-users@lists.schedmd.com
Subject: Re: [slurm-users] Suspend QOS help
Both jobs would be using the whole node same as below but with two nodes. I've
reduce
7:20 AM, Walls, Mitchell wrote:
> Hello,
>
> Hoping someone can shed some light on what is causing jobs to run on same
> nodes simultaneously rather than being actually suspended for the lower
> priority job? I can provide more info if someone can think of something to
> help
Hello,
Hoping someone can shed some light on what is causing jobs to run on same nodes
simultaneously rather than being actually suspended for the lower priority job?
I can provide more info if someone can think of something to help!
# Relevant config.
PreemptType=preempt/qos
PreemptMode=SUSPEN
For some reason it seems that DefaultQOS does not get set on the cluster level
when loading from file? Any ideas on why or if I have something wrong? I've
removed some output below for simplicity. As you can see below it is added when
doing sacctmgr modify cluster cluster set DefaultQOS='general
Do you see the uid in /sys/fs/cgroup? (i.e. find /sys/fs/cgroup -name
"*71953*"). If not that could point to cgroup config.
From: slurm-users on behalf of
Ratnasamy, Fritz
Sent: Friday, January 28, 2022 11:13 AM
To: Rémi Palancher; Slurm User Community
Are you using '#!/usr/bin/env python'? That could make a difference as an
example:
#!/usr/bin/env python3
#SBATCH --job-name=Python
from time import sleep
sleep(1)
print("done")
From: slurm-users on behalf of Walters,
Richard I ERDC-RDE-ITL-MS Contract
I am having an issue with limits not being enforced if I set
`AccountingStorageEnforce=limits,qos`, is this as expected? It allows
any user to submit jobs as well even if they do not have an
association. Seems to apply the default qos to every job but then
ignores associations and per user/account
